Environmental Informatics

(1)

Environmental Informatics

(2)

Environmental Informatics

(3)

Chapter 1. Environmental Informatics

Dr. Zoltán Utasi

This course is realized as a part of the TÁMOP-4.1.2.A/1-11/1-2011-0038 project.

1. Foreword

The textbook entitled Environmental Informatics is the continuation of Geoinformatic Applications the knowledge of which is required for the successful understanding and learning of this textbook.

Based on the curricula in the geography masters, the Geoinformatic Applications introduces the fundamentals of digital map construction from presenting the theoretical bases through the vectorising of paper-based maps to the construction of simple thematic maps. It is an outline basically as only the essential information was included due to extent limits.

The present textbook completes the previous one in several sections (e.g. regarding AutoCAD), however, several new applications are discussed as well (e.g. preparing digital elevation models). Moreover, those required working processes are overviewed that were regarded to be given (primarily the preparation of maps – starting from the mapping basics via the construction of raster maps to their georeferencing). Raster data analysis is discussed in more detail in accordance with the chapters of the textbook entitled “Analysing remote sensing data of satellites (Mika J. et al., 2011) written by the author of the present textbook.

It presents basically the software AutoCAD and Arc View outlined in previous textbooks together with other software when their application is limited.

2. Fundamentals of cartography

As those analysing maps do not always have the basic information to read maps this chapter gives an outline of the fundamentals of cartography. Basic terms associated with projections are explained, the various possibilities of surface visualization are listed, profiling systems of maps together with the mapping surveys in Hungary are discussed.

2.1. The fundamental question of mapping

A map in a very simplified sense is the view of the Earth surface from above constructed on the basis of certain mathematic-geometric rules showing the relief and land coverage elements (e.g. vegetation, drainage network, roads, buildings) as well. â€śReality” cannot be simply redrawn on a piece of paper, the most important factor

(8)

impeding this is the shape of the Earth. The fundamental problem of map construction is how a sphere could be unfolded without crease. In simple, by no means - just imagine a flattened beech ball, it cannot be completely smoothed. Therefore mapping aims not to achieve â€ścreaselessness” (crease called distortion henceforward) just to reduce it as much as possible or to shape it in order to achieve defined goals. Since most of us cannot see the Earth from the space, we have no image of the real shapes, we see only their distorted image on the maps leading us to false pictures numerous times. Figure 1.1 visualizes a well-known more-or-less spherical form - a human head - unfolded into a plane on the basis of various mapping procedures.

Figure 1.1: A human head in various projection types

If these forms seem to be distorted then compare the different illustrations of the American double continent (Figure 1.2).

(9)

3

(10)

Figure 1.2: Illustration of North and South America in different projections

None of these can be regarded good or wrong in absolute sense: all of them have different purposes. Differences are not so striking when smaller areas are depicted but great enough to make the work inaccurate of ignored.

Before we proceed a few basic mapping terms have to be defined:

• Base surface: the surface that is projected (e.g. Earth’s surface)

• Image surface: the surface, on which projection takes place (either planar or spatial - discussed in more detail later)

• Projection: the mathematic and/or geometric procedure on the basis of which the points of the base surface are matched with the points of the image surface unambiguously. (Unambiguous means in the mathematic sense that all points on the base surface are represented by one point on the image surface and vice versa.)

2.2. Shape of the Earth

2.2.1. Real shape of the Earth

The real shape of the Earth is given by the surface that can be defined clearly along the boundary between the solid and gaseous (terrestrial surface and the atmosphere) and between the solid and liquid phases (oceanic floor and the water mass). It is a completely irregular form, however, thus its transformation directly into plane would be difficult (Figure 1.3). Relief differences of the surface (20 km between the deepest point of the Mariana trench and the peak of the Mount Everest) are insignificant compared to the size of the Earth (r = 6378 km) therefore elevations are measured not to the centre of the Earth but to a - seemingly - explicit, clearly defined surface, the level of the seas and oceans.

Figure 1.3: Surfaces of the Earth

2.2.2. Theoretical shape of the Earth

When maps are constructed it is essential to know the accurate shape of the Earth as mathematic and/or geometric methods required to its projection can be determined only as a function of the shape. The shape of our planet is irregular, so called geoid that can be not or only with great difficulties transformed directly into plane therefore in a first step a simplification has to be made: surface elevation differences are ignored and we take as a basis a regular form that approaches reality as much as possible. If the planet was a body in rest with even mass distribution it would have a regular spherical shape due to gravity. None of the two conditions stand, however. Still assuming homogeneous distribution but taking the rotation of the planet into account a so called rotational ellipsoid would be obtained. This so called Newton type ellipsoid means a shape flattened at the poles and piking out along the Equator. Since this is only a theoretical shape (ignoring heterogeneity) the ellipsoid best approaching the geoid form cannot be determined directly, it can be deduced from numerous measurements. Therefore it was determined to be different at different time periods depending on the data and calculation methods available. The most important calculations are the following (Table 1.1):

Name year of calculation semi-major axis oblateness

Walbeck 1819 6 376 896 1/302.87

(11)

Bessel 1837 6 377 397.15 1/299.1518

Hayford 1924 6 378 388 1/297

Kraszovszkij 1940 6 378 245 1/298.3

IUGG 67 1967 6 378 160 1/298.247 167

WGS 84 1984 6 378 137 1/298.257 223 563

ETRS 1989 6 378 137 1/298.257 222 101

Table 1.1: Most important ellipsoids

The rotational ellipsoid is regarded as the base surface hereinafter. In the course of map construction the transformation of Earth’s surface points into plane is made in two steps. First, the points (P) of the topographic surface (real shape) to be projected are connected - in theory - to the centre of the Earth with the so called leading straight (h). The point obtained in this way (P0) on the base surface (rotational ellipsoid) is projected onto the plane in the second step (Figure 1.4).

Figure 1.4: Definitions of the shape of the Earth

Significance of the different base surfaces is that different co-ordinates are obtained if the same surface point is related to different ellipsoids (Table 1.2). Differences seem to be not significant but in reality these few arcsecond differences mean several hundreds of metres in reality.

Ellipsoid Geographical latitude Geographical longitude

International 47° 17Â´ 32.9Â´Â´ 19° 36Â´ 09.6Â´Â´

Kraszovszkij 47° 17Â´ 31.0Â´Â´ 19° 36Â´ 12.5Â´Â´

WGS-84 47° 17Â´ 28.8Â´Â´ 19° 36Â´ 06.7Â´Â´

Table 1.2: Geographical co-ordinates of a given point on the Earth’s surface related to different base surfaces (Papp-VĂˇry Ă�. 2007)

Due to the inhomogeneous mass distribution of the Earth, however, the so called geophysical level surface (the surface where gravity is equal) differs from the rotational ellipsoid creating the irregular shape called geoid. The difference of the two is measured as the geoid undulation (Figure 1.5).

(12)

Figure 1.5: Spatial differences in geoid undulation

2.2.3. Determination of height above sea level

In the course of mapping surveys in Hungary heights were measured to three different sea levels.

The first was the base level of the Adriatic Sea meaning the mean sea level measured (1875) on the scale of the Morto Sartorio in Trieste (Italy). This location was chosen because Trieste was the largest port of the empire.

Further base points were established in the interior of the country, however, only and the one in Nadap is found within the current borders.

In 1960 the socialist countries changed their base level to the Baltic Sea adjusting to the Kronstadt pier in the Finnish Bay east of Saint Petersburg. This level is 67.47 cm higher than that of the Adriatic Sea. (This is why the height of the Kékes peak is 1015 m in older maps and 1014 m on newer ones - and partly due to rounding as well.)

In 1994 Hungary accepted to change to the United European Levelling Network and thus to the base level at Amsterdam (only 14 cm below the Baltic one), however, its introduction is still in process.

2.3. Projections

Projection in mapping is the picturing of points, directions and figures of any rotational surface on another surface, plane or developable surface (Karsay F. 1962)

2.3.1. Classification of projections

Classification of projections can be made based on the following characteristics:

• distortion

• characteristics of mapping

2.3.1.1. Classification of projections based on distortion

When a spherical surface is illustrated in plane distortions appear. This includes the followings the value of which is given by the modulus:

- distance (length) - angle (direction) - area

(13)

Projections can be classified into four groups based on distortion:

- general distortion: all three quantities are distorted

- area retaining: proportional in area while distances and angles are distorted

- distance retaining (length retaining): area and angle are distorted and distance is proportional only in a certain direction, along lengths

- angle retaining: angles are real, area and distance are distorted

The grade of distortion of the three values is not independent from each other: decrease of one of them will increase the other two. Therefore those projections that are proportional in relation to one characteristic (area, distance and angle retaining projections) will distort the rest in high degree: they are applied generally to achieve a specific purpose (e.g. angle retaining projections in navigation). In the case of projections with general distortion all three values are increased only slightly therefore they are used in overview maps due to their values approaching reality.

2.3.1.2. Classification of projections based on the characteristics of mapping Parameters of mapping are given by the following four aspects:

Type of the image surface:

- 3D: mapping takes place onto a sphere - distortion is minimal, these are the globes

- 2D: mapping takes place onto a plane or onto a surface that can be projected onto a plane without crease - these are the maps. The image surface can be:

- plane sheet → plane projection

- cylinder surface → cylindrical projection - cone surface → cone projection

According to the relative location of the base and the image surfaces:

• hanging: they are distant away from each other with no touch or intersection

• tangent: they touch each other in a point or along a line

• intersecting: cross each other along one or two circles

According to the orientation of the base and the image surfaces, i.e. the angle between their axes:

• polar: they are parallel to each other, basically coincide with each other

• transversal: perpendicular to each other

• zenithal: they close an angle the value of which is between the previous two According to the method of mapping:

• Perspective: can be prepared mathematically (with calculation) and geometrically (with projection lines) as well. Two subtypes can be separated according to the characteristics of the projection lines:

• Parallel: projection lines are parallel to each other. If they reach the image surface in right angle then it is the orthographic type (orthogonal in other words), if they have any other angle with the image surface then the clinogonal type is obtained.

• Central: projection lines diverge from a centre of projection the location of which in relation to the base surface (Earth’s surface) separates the following two subtypes:

• external: outside

(14)

• stereographic: on the surface

• internal: inside, except for the centre

• central: in the centre

• Conventional (in other words: non perspective or mathematic, or sinusoidal): it can be prepared solely in a mathematic way.

2.3.2. Projection types

Design of the grid network is very characteristic for the given projection types, these are discussed in this subchapter.

2.3.2.1. Plane projections

Plane projections are generally circular. Their most important characteristic is that distortion is the same in the same distance from the point of origin of the projection, i.e. objects with similar distance in the projection have the same distance in reality as well.

In the case of the polar version, the pole is in the centre of the projection. Latitudes are circular, their distance depends on the type of projection. Longitudes are straights crossing each other with similar angles that are the same in reality. They are used generally to depict areas near the poles (Figure 1.6).

(15)

Figure 1.6: Stereographic polar tangent plane projection

In its transversal version the Equator is in the middle (in the form of a straight line), used generally for the illustration of one hemisphere (Figure 1.7).

(16)

Figure 1.7: Stereographic transversal tangent plane projection 2.3.2.2. Cylindrical projections

Perspective and conventional cylindrical projections have different shape. The most frequently used versions of cylindrical projections have polar orientation. A common characteristic of these is that the longitudes are straight parallel lines and the distance between them depends on the method of mapping. Latitudes are perpendicular to longitudes forming rectangles in this way. The pole is either a section with the length of the Equator or cannot be depicted. These characteristics also apply to certain conventional cylindrical projections (Figure 1.8).

(17)

Figure 1.8: Mercator's projection

Conventional cylindrical projections are variable in appearance. The projections most frequently used in military and civil topographic and cadastral mapping belong to them. The most important ones are the following:

Gauss-Krüger projection

The Gauss-Krüger projection is a transversally oriented, tangent, angle retaining cylindrical projection in which the Equator is uniform but depicts the surface in 6° wide belts (Figure 1.9). It was introduced to military mapping in Hungary at the beginning of the 1940s.

(18)

Figure 1.9: Situation of the ellipsoid and the cylinder in the case of the Gauss-Krüger projection and the developable dual angles

United National Projection (EOV)

The United National Projection is a zenithal oriented, intersectional, angle retaining cylindrical projection and it is the current official projection of civil mapping in Hungary (Figure 1.10).

Figure 1.10: Location of the EOV (Varga J. 2005) UTM (Universal Transverse Mercator)

The UTM projection is a transversal oriented, intersectional, angle retaining cylindrical projection (Figure 1.11).

(19)

Figure 1.11: Location of the UTM (Varga J.2005) Goode projection

The Good projection is a complex projection type. The border between the Mollweide and Sanson projections, its components is the 40° latitude. All continents have their separate straight starting meridian thus the projection is divided into 5 or 6 parts. As it is area retaining and thanks to its dissection its distance and direction distortion are small it is used frequently to depict terrestrial features on global scale (Figure 1.12).

Figure 1.12: Goode projection (Wikimedia Commons) 2.3.2.3. Cone projections

Cone projections have pie shape (all of the perspective type and the conventional types are similar as well) in the case of the most frequently used polar orientation the pole is located in the centre of the projection. It is easy to mistake it for the polar plane projections if only a part of it can be seen as the latitudes are circular while longitudes are straight. They intersect each other with similar but - different from plane projections - smaller angles than in reality (Figure 1.13).

(20)

They are very frequently used for maps with large scale (e.g. in atlases), generally central and low latitude areas are shown in them.

Figure 1.13: Central polar tangent cone projection

Following the overview of projections it is necessary to define the term scale (M) of maps. According to the popular but inaccurate definition: M = distance on map / real distance. As no maps are distance retaining in every direction (because it should be angle and area retaining at the same time) distances measured similar on the map will differ in reality. Based on this, the accurate definition is: M = distance on map / distance on projection.

2.4. Segmenting

Segments of topographic and cadastral maps (the area displayed) are not adjusted to natural or political boundaries (as in the case of most maps used in public) but the system is composed of map segments ordered in a well defined way. Requirements of segmenting:

• Homogeneity: base surface, base level, projection, co-ordinate system, legend, contents of maps are the same in all segments.

• Spatial continuity: joined without overlapping or gap, maps fill the entire area.

(21)

• Up-to-date: corrections, revisions are required at certain intervals.

• Back-to-back state: members with different scale in the system are based on each other in the framework system, i.e. segments with greater scale form the mosaic of different size of the maps with smaller scale.

Identification of the segments (geographical definition) can take place in two ways: on the basis of a segment number or segment name. Giving a name to a segment follows the following rules:

• It may receive its name after the most significant settlement located in the segment or after any significant object (e.g. mountain peak) if no settlements are found in the segment.

• Generally the name of two segments cannot be the same in a given scale (except for EOTR). If a settlement is divided on several segments then some kind of an indexing is required, like a segment called Cered and another segment called Cered (Szekrényke-puszta) are present.

Naming the segments is simpler, however, less accurate as it may require significant topographic knowledge (e.g. it is difficult to determine the most significant denominative settlement for a user who does not know the area).

Segment numbers define the location of the segment with the combination of letters and numbers according to a given system. This is more explicit although more difficult to use from certain aspects than the segment name (detailed knowledge of the system is necessary). Currently used and older segmenting systems in Hungary are overviewed in the followings:

2.4.1. Segmenting system of military survey III (Austro-Hungarian Monarchy)

The military surveys of the Austrian Empire and then the Austro-Hungarian Monarchy are discussed because these maps are still used today (e.g. studying changes of the land).

The starting meridian of the survey was Ferro and the frame of the segments is given by the grid network (grid trapezes). Picturing was made on the following scales:

- The scale of the initial segment is 1:200000, its size is 1° x 1°. This has no separate number.

- The first deduced scale is 1:75000 obtained with the division of the initial segment into 2x4 parts. Numbers are obtained by the determination of the row and column (e.g. 1242), direction of numbering is east and south (Figure 1.14).

- The second deduced scale is 1:50000, obtained by the N-S bisection of the previous segment (E, W). Its numbering: row-column-page (e.g. 1242 Ny).

(22)

Figure 1.14: Segmenting of the military maps of the Austro-Hungarian Monarchy

2.4.2. Segmenting of the International World Maps (Gauss-Krüger projection system)

The International World Map covers the entire Earth the starting meridian of which crosses Greenwich, its base surface is the IUGG/67 or the Kraszovszkij, its projection is the Gauss-Krüger (transversal, tangent, angle retaining cylindrical projection). The frame of the segments is given by the grid network (grid trapezes).

- Scale of the initial segment is 1:1000000, its size: 6° x 4°. Its numbering is based on (hemisphere)-row-column (e.g. N-L-36). Direction of numbering:

- Row: increases to the north and south from the Equator (A-Z) - Column: increases to the east from 180° (1-60)

- In the case of the deduced scales the frame numbering of the first three (1:500000, 1:250000, 1:100000) is retained but the inner division is changed (Figure 1.15).

- In the case of scales less than 1:100000 quartering is made in every steps, the location of these quarters is marked with retaining the frame number of the 1:100000 scale (Figures 1.15 and 1.16).

(23)

Figure 1.15: Segmenting system of the International World Map

(24)

Figure 1.16: Overview of the segments of the Gauss-Krüger maps with the scale of 1:100000 and the further division of the segments (Papp-VĂˇry Ă�. 2007)

2.4.3. Segmenting of the United National Mapping Network (EOTR)

The EOTR is used exclusively in Hungary, its projection is the EOV (United National Projection) that is a zenithal, intersectional, angle retaining cylindrical projection. Its starting meridian crosses Greenwich, its base surface is the IUGG/67. Its segments are framed by not the grid network but a kilometre network the starting point of which is the Gellért Hill, its numbering is shifted 651 km and 200 km to the west and south from the starting point respectively.

- The scale of the initial segment is 1:100000 the size of which is 48x32 km. The segment number includes the number of the row (0 - 10) and the column (0 - 11) (e.g. 03, 56, 109, 910), direction of numbering is east and north (Figure 1.17). Two important differences compared to the rest of the systems have to be noted:

- Numbering of rows and columns may start with 0 as well.

- In case the number of both the row and the column is of one digit the segment number is composed of two digits but if one number is of two digits the segment number is of three that makes it hard to be interpreted sometimes. For example, 811 means row 8 and column 11, however, it could also mean row 81 and column 1:

in this case the user have to know that the system is composed of only 11 rows!

- Topographic deduced scales (from 1:50000 to 1:10000) are obtained by the repeated quartering of the initial segment (bisection in both directions). The segment number is obtained by the extension of the frame number with the number of the given quarter (e.g. 56-444) (Figure 1.17).

- In the cadastral scale (under 1:10000) segments are still obtained by quartering and the segment number is composed according to the frame number - topography deduced scale - cadastral scale principle (e.g. 56-444- 444) (Figure 1.17).

(25)

Figure 1.17: Overview of the EOTR segments with the scale of 1:100000 and the further division of a given segment (Papp-VĂˇry Ă�. 2007)

2.4.4. National segmenting

Using the International World Maps discussed in chapter 1.4.2 is difficult in Hungary as the country is depicted in four 1:1000000 scale segments and the determination of the continuing segments is more complex at the corners (Figure 1.16). Therefore the segments are renumbered with retaining the parameters of the original map.

This numbering is only valid for the area of the country.

The initial scale is 1:100000 (like in the case of the EOTR as smaller scales are pointless), these segments are numbered with the row-column logic (like in the case of the military Survey III or the maps of the EOTR system), i.e. the segment number is composed of the row (1-9) and column (1-14) values (e.g. 113). Segment numbers have always three digits, the first representing the row value and the second and third indicate the column value - if the latter is smaller than ten it will start with zero (e.g. 506 represents row 5 and column 6).

Deduced scales are obtained by quartering the initial frame, numbering is made according to the same logic as in the case of the EOTR (chapter 1.4.3).

2.4.5. UTM (Universal Transversal Mercator) segmenting (NATO system)

The UTM segmenting covers the entire World the starting meridian of which crosses Greenwich, its base surface is the WGS-84 and its projection is the universal Mercator projection (transversal, tangent, angle retaining cylindrical projection). Frame of the segments is given by the grid network (grid trapezes).

- Scale of the initial segment is 1:1000000, its size is 6° x 8°. Its numbering is based on the row-column logic (e.g. T-33). Direction of numbering (Figure 1.18):

(26)

- Row: Increasing towards north from the South Pole (A-Z).

- Column: Increasing towards east from 180°.

Segments and numbering are differ a little at the poles.

Figure 1.18: Segmenting of the UTM projection

2.5. Content of maps

Following the determination of the co-ordinates and the location of the projection of the points the elements of the surface are depicted. These elements can be the following:

• Discrete elements: they have boundary lines. Their depiction is simple as the boundary line has to be depicted, measurement points have to be market on it (several or only a few in a unit section depending on curvature). In numerous cases, however, the boundary line of neighbouring well definable elements is not sharp, they have overlapping belts (so called fuzzy). For example in the case of detailed scale the boundary between a forest and a grassland is hard to be determined: the edge of the leafage or the trunk of the trees represent the boundary?

• Continuous elements: elements with continuous distribution. Their depiction is more complex as the constructor of the map has to create value degrees with the help for example of isogons (lines connecting the points of identical values). Such lines â€śdo not exist” in reality. For example meteorological-climatological maps (precipitation, temperature, etc.), ethnic maps and the list could be extended. Relief also belongs to this category.

Maps can be classified as follows based on their content:

• Topographic maps: they show as many elements of reality as possible trying to achieve completeness (including for example relief, surface coverage, anthropogenic elements).

• Thematic maps: highlights only one or some elements of reality (e.g. temperature map) or it contains deduced data (e.g. population density map where calculation is made on the basis of the ratio of area and population).

• A special type of thematic maps is the cadastral map that contains basically the boundary of areas (parcels) and the points necessary for identification (e.g. triangulation points).

(27)

2.6. Relief illustration

Most important task of relief illustration is to present the parameters of relief accurately. Specifics of the surface can be illustrated in two ways fundamentally: considering the height of surface or the angle of slopes. Generally different illustration methods focus on only one of these, the other cannot be or only hardly illustrated.

2.6.1. Relief illustration by stripes

Relief illustration by stripes was established in order to depict the quantitative characteristics of the surface accurately. The system founded by Lehmann in Prussia in the 18^th century was aimed for military purposes: the aims were the accurate depiction of slope steepness and good arrangement.

Principle of illustration is striping parallel to the slope direction and with thickness proportional to slope steepness (Figure 1. 19). Based on these, gentle sloping areas remain white while steeper areas look darker due to dense striping (1.1. animation). The result is accurate and easy to read (even for a non-professional).

Its disadvantages include the problem of depicting height above sea level: based on the principle, this cannot be determined (only sloping) therefore numerous height values help height orientation. These show generally the values of known places (peaks, road-junctions, important objects, etc.). Depicting land coverage also present a problem: due to dense striping map symbols almost shade into their surroundings or mask them out. The same causes make the depiction of plant cover also problematic. This latter was partly solved by applying colour technology (Figure 1.20).

In summary the transformation of this method of illustration into vector digital files is very complicated (since digital elevation models require primarily the height values and sloping values are calculated from this by interpolation). Determination of vegetation boundaries is also problematic (as there is no boundary between the map symbols!).

This type of illustration was replaced by relief illustration by contour lines (chapter 1.6.4) at the beginning of the 20^th century but it was used for a long time. Figure 1.20 shows a partly updated segment: revision to contour line illustration was made north of the temporary country borderline represented by the red line while south of it the striping method is used.

1.1. animation: Detailes of relief illustration by stripes

(28)

Figure 1.19: Relief illustration by stripes

Figure 1.20: Map segment partly updated at the time of World War II

(29)

2.6.2. Topographic shading

The principle of topographic shading is that light rays arriving vertically light the surface that is regarded to be homogeneous proportionally to slope steepness therefore steeper areas seem to be darker while more gentle slopes lighter (Figure 1.21).

This makes an impression similar to striping but here filling is performed by homogeneous colouring. Its advantages and disadvantages are also very much the same as those of striping as it is simple to read, spectacular and slope steepness can be determined well, however, height is not depicted either (only height numbers at the most) and coverage can be depicted hardly as well.

It is rarely used alone, mostly combined with height colouring. It is very difficult to vectorize this type.

Figure 1.21: Topographic shading

2.6.3. Height colouring

Height colouring marks the surface with different colours for height classes (Figure 1.22). Boundary of the height zones is given by the line connecting the points of identical heights above sea level (asl.). Interval of the zones is not similar, it increases from the sea level (downwards as well). The reasons for this are on the one hand that one unit of height difference means greater change in lower heights and on the other hand that the primary area of human activity is found in lower heights, therefore their detailed presentation is justified.

When selecting the colours the principle of progressivity has to be considered, heights become darker away from sea level (except for the dark green colour of dale-land) together with tradition on the basis of which lowlands are green, hills have yellowish brown and mountains brown colour.

The number of zones (categories) is low generally (8-10) as the difference between the colour of neighbouring zones would be too small making their reading difficult.

The method is generally applied together with a special type of topographic shading. Its advantage includes the easy determination of height asl. - within an interval - and it is easy to read and spectacular. Its disadvantages are the determination of slope angle is not possible and the difficulty in depicting land cover (with map symbols at the most).

It is used frequently in the case of overview maps (e.g. atlases) but it cannot be used for detailed presentation (due to the limited number of categories). It can be vectorized, however, its application is limited (due to the small number of height zones).

(30)

Figure 1.22: Depicting relief with height colouring

2.6.4. Depicting relief with contour lines

Relief presentation with contour lines is still the most widespread method in detailed mapping as the elements of the Earth’s surface are illustrated in greatest ratio with this method. Contours represent the basis of depiction, contour lines connect the points of the surface that have identical heights above sea level (Figure 1. 23).

(31)

Figure 1.23: Relief depiction by contour lines

Numbering of the contour lines always starts from seal level. It has several types (Figure 1.24):

• Base contour line: its colour ranges from yellow to brown (occasionally black), it is a continuous thin line.

Their height difference is the same. They cannot be broken, only in a few special cases (Figure 1.25). These are the following:

• Deep linear form

• Extended steep slope

• Another map element masks it

• Main contour line: base contours are thickened at certain - similar - height intervals in order to make readability simpler. Generally every fifth (e.g. in the case of 5 m base contour intervals the interval between the main contours is 25 m), however, if this would be not a round value main contours can have smaller intervals (e.g. in the case of base contour interval of 2.5 m every fourth is the main contour with intervals of 10 m). Numbering of main contours also starts from the sea level.

• Auxiliary contour lines: when orography requires (mainly in the case of small height differences in greater areas) the height interval of base contours can be halved by a bisecting auxiliary contour line (generally long dashed lines with colour and thickness similar to the base contours) or it can be quartered by a quartering auxiliary contour line (generally shorter dashed with colour and thickness similar to base contours).

(32)

Figure 1.24: Contour line types

Figure 1.25: Types of contour line breaking

The interval between base contour lines depends on the scale. It is sensible that with the decrease of resolution (decrease of scale) contours would become closer to each other and they would merge together at a certain

(33)

resolution therefore the interval between the base contours has to be increased. For example on International World Maps the base contour intervals are 2.5 m, 5 m and 10 m on the scales of 1:10000, 1:25000 and 1:50000 respectively.

Base contour intervals - and thus main contour intervals as well - are similar on most of the topographic maps therefore height determination is relatively simple. In contrast in the case of the EOTR the base contour interval depends on the dissection of relief. Base contour interval is 5 m, however, in extended, gentle sloping areas it changes into base contour intervals of 1 m (value of the bisecting auxiliary contours is 0.5 m). At the location of change the base contour is thickened and its value is written - increased attention, however, is essential to recognise it (in Figure 1.25 this change is seen in the lower part). In summary, this solution increases topographic accuracy but in the meantime it also increases the significance of map reading (Figure 1.26).

(34)

Figure 1.25: Change of base levels in EOTR maps (1.)

(35)

Figure 1.26: Changes in base contour intervals in EOTR maps (2.)

Relief depiction by contour lines combines nearly all of the advantages of the methods discussed above (striping, topographic shading, height colouring):

• Height above sea level can be determined accurately

• Slope angle can be determined accurately

• Land cover can be shown clearly

Unfortunately the "price" of accuracy and completeness is readability: its usage requires practice and experience.

2.7. Land cover and geographical names

Relief is only one - doubtless important - element of reality, various elements are found on it that are called land cover. This includes both natural and artificial elements, biotic and abiotic environment as well.

Showing vegetation depends on the purpose of the map. In certain cases the primary aim is to present the morphology: on overview maps (e.g. in atlases) when height colouring is used in order to illustrate surface specifics as much as possible the presentation of vegetation is very limited (with map symbols at the most). In the case of detailed maps it is a fundamental requirement to show vegetation. Fullness of representation, however, is variable: in contrast to the few categories of tourist maps available for the public (e.g. forest - bushes - area with low vegetation) survey maps can depict several tens of categories by colours and several dozens of vegetation types within them using map symbols. Generally the hue of colour corresponds to the height of the vegetation: from the dark green of forests through the middle green of bushes to the white of cultivated lands or grasslands. Exception is the orienteer map where hue represents the penetrability of the vegetation: darkest ones are most difficult and white areas are least difficult to overcome.

Illustration of anthropogenic elements is also variable. Map symbols of linear elements (e.g. transport network) generally use colour, line thickness and type to show the specifics. Map symbols of buildings indicate their building material (e.g. type of brickwork, fire resistance) and their function.

Names illustrate the specifics of the named object by their letter type, style, placing, colour, size.

• Names of relief elements are generally parallel to the strike of the landforms, the letter size is proportional to the real extent of the form.

(36)

• In the case of rivers their names are oriented parallel to the rivers and their direction points - generally - into flow direction. Lake names are either parallel to the strike of the lake bed or horizontal. The names are mostly blue, however, other colours are also used.

• Name of anthropogenic elements is generally horizontal their size and letter type are proportional to the size and significance of the given element (e.g. in the case of settlements letter size is increased and letter style is changed with the number of inhabitants and with the place of the settlement in the public administration hierarchy).

2.8. Cadastral maps

A special, however, widespread type of thematic maps is the so called cadastral (land surveying) map. The primary aim of this is to register property boundaries. Its content is very poor compared to topographic maps:

only the boundary of areas, buildings and points necessary for orientation (e.g. triangulation points are shown).

On certain versions contour lines also appear.

The projection system of the cadastral maps used in Hungary today is the EOV (chapter 1.3.2.2), its segmenting is the EOTR (chapter 1.4.3).

Cadastral maps of the non-built-up areas of settlements (Figure 1.27) show only the legally non-built-up areas while the built-up area of settlements remains empty (Figure 1.27). It contains not only the land register reference of the areas but the cultivation types as well.

Cadastral maps of the built-up areas of settlements (Figure 1.28) contain the profile of buildings as well besides the borderline of plots. Apart from land register reference they may contain street-numbers as well.

Figure 1.27: Cadastral map of non-built-up areas of settlements (section)

(37)

Figure 1.28: Cadastral map of built-up areas of settlements (section)

2.9. Surveys in Hungary

2.9.1. Military surveys

The detailed systematic topographic mapping in the Austrian Empire - Hungary was part of it at that time - started in the middle of the 18^th century with military purposes. Several survey series were completed by today, however, only military surveys I - III were complete as they covered the entire country in a - relatively - short period of time (Table 1.3). From the beginning of the 20^th century only the update of the existing files was carried out, comprehensive operations were not performed.

I II III

Period of survey in the Hungarian Kingdom (without Transylvania, Croatia, Slavonia and the defended frontier region)

1766-1785 1806-1869 1872-1884

Scale of the complete survey 1:28 800 1:28 800 1:25 000

Size of one segment 24x16 inch

(63x42 cm)

24x16 inch

(20x20 inch outside Hungary)

76x55 cm

Size of the area depicted in one segment, km² 216 216

(230 outside Hungary)

261

(38)

Number of segments 963 1078 1353

Unit fathom fathom metre

Availability secret survey maps are secret, deduced

maps are open

open

Table 1.3.: Overview of military surveys in Hungary (Papp-VĂˇry Ă�.)

2.9.1.1. First military survey of Hungary (Josephinian map) (1766 – 1785)

The first military topographic mapping that depicts the entire historical Hungary below settlement level. The scale of 1:18800 of the maps constructed in fathom system is explaneed by that 1 inch on the map equals 40 fathoms in reality (1 Vienna fathom (1.8964838 m) = 6 feet, 1 foot = 12 inches, i.e. 40x12x6 = 2880). Relief is shown using stripes, the maps also contain hydrological elements (rivers, lakes, roads, settlement and geographical object names and the cultivation type of agricultural areas just for information. The survey was not united in the empire so it is composed of separate sections and the location of the survey was determined by the risk of threat (this is why Silesia was completed first) (Figure 1.29). Accuracy of the work without a projection system is satisfactory in general within the individual segments, however, at the connections significant differences (up to 200 metres) are also found. No height measurements of levelling were performed in the course of the survey and the maps contain no height data. Georeferencing of the maps is very complicated due to its inaccuracy.

Figure 1.29: The Josephinian military survey (Arcanum)

2.9.1.2. Second military survey in Hungary (Franciscan map) (1806 – 1869)

The second survey was performed on the basis of the Josephinian maps, they were actualized and corrected (Figure 1.30). The survey was ordered by Francis I in 1806 as the effect of the war with Napoleon. The aim was to compile maps where the survey and depiction of the empire are performed according to united principles.

Cassini’s projection and segmenting system was used with slight modifications. This is as a matter of fact, the Cassini-Soldner projection due to its inaccuracy it is frequently referred to as the system without projection.

(39)

The second triangulation base point network of Hungary was established in this time period. This served as the basis for both military topography and civil cadastral mapping. Out of the ten systems of the Empire three covered the area of Hungary (Figure 1.31). Height measurements were also carried out.

The map is extremely difficult to read at places due to the method of the dark tone relief depiction. Its georeference, however, is easier - due to its higher accuracy - than that of the previous one (Figure 1.32).

Figure 1.30: The Franciscan military survey (Arcanum)

(40)

Figure 1.31: Systems without projection in Hungary (Varga J., 2005)

(41)

Figure 1.32: The area of Battonya on the maps of the first and second military surveys

2.9.1.3. Third military mapping in Hungary (Franciscan-Josephine map) (1872 – 1884)

The maps were constructed following the introduction of the metric system therefore their scale is different from the 1:28800 applied before (in the case of both the Josephinian and Franciscan maps). The introduction of the scale 1:25000 into military topographic mapping turned out to be long-term, still applied today. The second

(42)

triangulation base point network completed between 1808 and 1861 was used with appropriate increase of base point numbers. Its method of depiction is different from the previous two surveys as graphic data related to relief and drainage network have smaller significance. In the meantime, the most important road names and object symbols occur on the maps together with height data.

2.9.1.4. Recent topographic maps

Military and civil mapping after harmonizing until World War I, separated sharply following World War II. The aim was to establish two completely independent systems that are not compatible to each other.

At the beginning of the 1940s military high command decided to introduce the Bessel ellipsoid based Gauss- Krüger projection in which topographic maps were published only in German prior to World War II. Following the war the system was changed to Kraszovszkij ellipsoid. Survey maps were prepared on the scale of 1:25000 with international segmenting (chapter 1.4.2) and of course areas over the border were also shown in the segments.

Civil mapping also established its own system. This was also based on the Kraszovszkij ellipsoid but it applied stereographic projection and a segmenting different from the military one (chapters 1.4.3 and 1.4.4). Areas over the border were not shown. Following the regime change in 1990 the United National Mapping System used exclusively in Hungary (and its projection, the EOV) became the official system and contents of maps were started to be transformed into it. At present the transformation is complete but map contents have not been updated therefore it is possible that the â€śmost recent” EOTR map of an area was constructed in the 1990s but it is based on data taken from 20-30 years older maps! Areas over the border are still white patches.

The European Union brought changes in mapping as well: the new ETRS 1989 LAEA standard will (may) become the official mapping system of the member states in the future.

2.9.2. Cadastral maps

The first cadastral survey in Hungary was ordered by Joseph II and carried out between 1786 and 1789. It was slow due to the opposition from the nobles as they feared the data would be used for their taxation as well. Later the survey came to a halt and the already prepared maps were destroyed after the death of the king.

The second cadastral survey was carried out between 1856 and 1894 on the scale of 1:2880 in the fathom system. The reason for the 1:2880 scale is that 1 inch on the map equals 40 fathoms in reality and 1 Vienna fathom (1.8964838 m) = 6 feet, 1 foot = 12 inches, i.e. 40x12x6 = 2880.

The third cadastral survey was performed between 1900 and 1938 still in fathom system and on the scale of 1:2880 (1:1440 in towns).

Following World War II the change to the metric system took place in civil mapping as well. The new cadastral maps are constructed in the EOV system on the scales of 1:8000, 1:4000 and 1:2000 (see chapter 1.4.3).

Controlling questions Self controlling questions:

What projection systems are used in Hungary as well?

What segmenting systems are used also in Hungary?

What advantages and disadvantages does contour line relief depiction have?

Test:

Which name means base surface (ellipsoid)?

a, Gauss-Krüger b, Kraszovszkij c, Goode

(43)

What relief depiction method was used in the course of military mapping I and II?

a, Striping b, Contour lines

What is the meaning of cadastral map?

a, Military topographic map b, Base map for land registering

3. Preparation of map files

The chapter discusses the practical issues of transforming paper based maps to raster digital format, the frequent errors of vector files and the basics of georeferencing raster map files.

3.1. Data sources

Object data can be acquired from several sources that can be classified into two groups regarding originality:

primary and secondary data acquisition methods.

Primary data acquisition methods include all methods in which determination of the parameters of the objects and the Earth surface is direct. Sources can be the following:

• traditional land geodesy

• Satellite positioning

• Remote sensing

• Photogrammetrics

In the course of secondary data acquisition maps composed on the basis of field survey are conversed (reworked, modified) - occasionally at several times - using the following methods:

• Manual scanning already existing maps (raster)

• Manual digitizing of existing paper based maps (vector)

• Takeover of existing digital maps (raster, vector)

• Data derived from existing digital data (e.g. runoff, slope orientation and inflection maps, raster, vector from digital elevation models)

Secondary data acquisition is always accompanied by data loss (regarding accuracy and content completeness as well) the minimizing of which is among the most important aims. With the number of conversions the problems increase. For example the repeated processing of a scanned, modified and printed map in a publication shall be avoided, the original source should be used.

3.2. Transforming paper maps into raster format

One of the most frequent preparations in our work is the digitizing of the paper maps resulting in a raster file that will be the basis for further work processes (e.g. vectorizing).

3.2.1. Scanning paper maps

The sub-chapter discusses the problems of scanning and their possible solutions.

3.2.1.1. Preparation of paper maps

(44)

Paper maps require careful handling before scanning. Most frequent problems and their possible solutions are the followings:

• The map is folded or even torn: due to improper storage the paper is not flat completely and this may result in distortion. The most frequent types and reasons of this:

• The map was stored rolled up in a cylindrical form: rolling it u pin the opposite direction can flatten the map to a certain degree.

• Bent along an edge: this can be reduced by preliminary pressing at best and it is almost impossible to eliminate completely. If multiple bending took place with crossing edges significant folds can appear in the cross-points the smoothing of which is very difficult.

• Irregularly folded: only preliminary pressing may help, however, perfect results cannot be expected.

• Torn map sheet: if possible (e.g. not a piece from a museum) stick the pieces together, however, perfect joining cannot be expected.

• Up-warped due to moisture: curliness can be reduced by pressing at best.

Please, note that similar scanning of a given map sheet cannot be performed twice (as some kind of folding appears even if the maps handled with maximum care).

3.2.1.2. Original scanning of map sheets

In the course of scanning maps the size of the sheets presents the greatest problem as using the most frequent A/4 sized scanners the maps of average size can be scanned only in parts (in the case of a normal topographic map this can be 8-10 parts) and at the subsequent adjustment of the parts errors arise almost every time (for reasons and possibilities see later). Using a larger scanner can be a solution, however, this increases the costs as well: even the A/3 size (more expensive by a magnitude) is not enough and the A/0 size scanning means two magnitudes greater cost.

Besides the size of scanning the rest of the parameters are important, however, not as mush as in the case of scanning photos for example: resolution is adequate until a maximum 600 DPI, colour depth is not important (as maps contain less colour compared to photos). Software modification of the scanned image could be important, like reducing granularity or - if necessary - increasing contrast. In the case of publications with less quality granularity caused by the printing technology is visible at high resolution and increases the size of the file unnecessarily. The following figure series (Figures 2.1 - 2.4) show the same map segment modified by various techniques. Readability differences among the versions are visible: since increasing the contrast increases file size as well we have to find the optimum solution. In practice the solution shown in Figure 2.3 is the best generally.

(45)

Figure 2.1: Map part scanned with default settings

Figure 2.2: Map part scanned with increased contrast

Figure 2.3: Map scanned with reduced granularity

(46)

Figure 2.4: Map part scanned with reduced granularity and increased contrast

In the case of uncompressed format (e.g. bmp format) there is no significant difference between the sizes of the files. However, in the case of compressed format the difference can be significant: considering the sample files the increased contrast version (Figure 2.2) is one and a half times while the granularity reduced versions (Figures 2.3 and 2.4) are half as great as the default one (Figure 2.1)

3.2.1.3. Repeated scanning of map sheets

It is possible that the original scanning produces a file of poor quality therefore a repeated scanning of the map would seem to be sensible. This presents a problem if the original version has already been processed (e.g.

vectorizing was performed on its basis). In this case the replacement of the map is very difficult: it is almost impossible to crop exactly the same part therefore even if the positioning of the map was according to known points the new one will not cover the original one completely (e.g. due to the folding mentioned before). Due to the above it is possible that though a part of the vectorizing had been completed based on the map to be replaced then this map is replaced by the new version but its content (lines, points) do not fit exactly to the vectorized objects. In conclusion repeated scanning is only recommended if no processing has been performed.

Figure 2.5 shows digitizing on the original base map: vectorized lines overly exactly the lines of the map.

Figure 2.6 illustrates the same vectorized elements placed onto a re-scanned and inserted map: some lines overly exactly the original lines while others (especially in the central part of the map) do not.

(47)

Figure 2.5: Vectorizing on the original map

(48)

Figure 2.6: Original vectorized content on a re-scanned map

3.2.2. Joining raster (digital) content

In case the map cannot be scanned as a whole there are several possibilities to join the parts.

• Effective merging of the raster files that can be performed either:

• manually or

• automatically

• Placing each image beside each other in a vectorizing software 3.2.2.1. Effective merging of raster files manually

Manual merging means in practice that the size of the first part is increased in order to place the rest of the pieces beside it - using the measure taken by the eye. This procedure requires great care as the pieces were not scanned in exactly the same position: it is virtually impossible to place the map on the scanner at exactly the same position, it will rotate towards the left or right compared to the previous position. In the course of the merging the parts have to be rotated the grade of which has to be determined very accurately. (At the same time rotation always involves the deterioration of image quality!) Even at greatest care the maps will not fit perfectly to each other: the map folds a bit differently at each time it is placed onto the scanner and this cannot be eliminated. According to the experience of the author, however, if the map is handled with care (spare them from folding - e.g. they are not folded, only rolled up to a cylindrical form) this type of joining yields reasonable results. Discussing image editor software is not included in the aims of this textbook (due to the limited extent) but any software of reasonable quality is capable of the task (e.g. Paint Shop Pro, Adobe Photoshop). Figure 2.7 shows the limits of manual merging: on the left side of the figure fitting is accurate while on the right side certain lines are drifted.

(49)

Figure 2.7: Errors of completely manual merging (Paint Shop Pro) 3.2.2.2. Effective merging of raster files automatically

For this any general image editor software that supports the making of panorama images or any special software designed especially for the purpose (e.g. Panorama Factory) can be used. The procedure is based on that the software merges the images on the basis of finding identical points on the overlapping parts of the images. In the course of the rubber-sheet procedure the images are distorted until they fit perfectly. This procedure in theory able to eliminate errors not only of different positions but of folding as well, however, according to the experience of the author the overlapping zone is frequently ghost image. This not striking in the case of photos - for the merging of which the software is designed - but in the case of the linear elements of maps this is confusing. More detailed discussion of this procedure would exceed the limits and the fundamental aims of this textbook. Figure 2.8 shows the automatic fitting of the area presented in the previous sub-chapter (Figure 2.7):

connecting points are more accurate but the image is much noisier, less clear.

Figure 2.8: Automatic fitting (manual document merging option, Panorama Factory) 3.2.2.3. Fitting image sin a vectorizing software

In simple cases merging can take place virtually as well, i.e. the parts are not merged in reality they are only placed next to each other at the appropriate position in the given vectorizing software. In the case of AutoCAD discussed in the textbook the image parts are measured and positioned separately (see chapter 10.3 of Geoinformatic Applications). This method may cause problems in the overlapping parts as these parts do not overlap each other accurately (due to the reasons discussed in the chapter on scanning, e.g. due to the folding of the paper) therefore it is important which parts the vectorizing were based on.

3.2.3. Determination of the segment of maps

Either a map was scanned in one or in parts we have to decide whether the frames are needed or not. There are arguments pro and contra so decision shall be made following the careful analysis of the exact task.

Arguments for retaining the frame:

• The frame of a map contains numerous important information the more important of which are:

• Segment name and number: less important as they can be displayed in the file name.

• Neighbouring segments: also less important as they can be looked in an overview map at any time.

• Values of the grid (degree, kilometre): very important as they are required for the appropriate positioning of the maps (georeferencing). Very few or no data related to this can be found in the inner parts of the map.

• Complementary maps (overview maps, data survey maps, etc.) may contain numerous useful information.

Arguments against retaining the frame:

(50)

• Increases file size.

• Wide overlapping areas will occur when neighbouring maps are joined and these cover each other. This can be disturbing in the course of vectorizing (can be eliminated by changing their order in AutoCAD).

• Results in inaccuracy in the case of georeferenced maps. This can be explained by that georeferencing is based on t he co-ordinates of the corners related to the map. When the map has a frame co-ordinates are shifted and cannot be given accurately.

Figure 2.9 shows the vectorizing of the inserted map first: lines run exactly until the frame of the map while the frame is visible on the map itself (with the values of the grid and kilometre network). Figure 2.10 presents the neighbouring map - from the east - on top: lines digitized from the other map and running until the edge of the neighbouring map are clearly visible. The lines could be continued but the frame of the overlying map masks the western margin of the map below.

If we decide to cut the framed we have to consider that the sides of the maps will not be exactly vertical and horizontal (as in most segmenting systems the grid network gives the border of the segments, except for the EOTR system) therefore a belt without map content will remain.

Figure 2.9: Neighbouring segment underneath

Environmental Informatics